*This session is facilitated by Brian Obilo, Maurine Basoi *
About this session
In our interactive session, we will introduce our audience to biases in AI systems. The audience will get an opportunity to shape the discussions on bias and they’ll be given the time to write down and share possible ways in which bias may manifest in AI. We will then take them through the process of identifying areas where bias may creep into an AI system. The audience will then identify what we can control and will then give a practical demonstration of how you can use a second layer of AI to counter bias on a sample AI system.
It’s important for our audience to understand that AI is only as healthy as the data it runs on and most importantly, the humans that interact with it. Therefore, it is vital that they will see (firsthand) the importance of rooting out bias at every stage.
Goals of this session
The main goal of our session is to sensitize our audience on how to discover and effectively correct biases in AI systems. AI systems are vulnerable to various forms of bias such as unconscious bias, explicit bias and ignorance in some instances.
In a nutshell, our audience will get the opportunity to understand how they can correct bias and improve the fidelity of their AI systems. Our presentation targets not only different age groups, but people of different skill levels as well.